Previously we talked about adding to a dataset a whole set with the fit line, but I have a better idea. Basically, almost all the standard fits (Gaussian, Lorentzian, etc.) can be characterized by a few basic values (mean, std for Gaussian, etc). Instead of saving a whole set of data, we can just store these as parameters of the dataset in the hdf5 file, under names that make sense (maybe fit_gaus_mean or something like that). We then can have a checkbox to turn the plotting of the fit on or off. It is probably easiest to run the fitting algorithms as a plugin.
Previously we talked about adding to a dataset a whole set with the fit line, but I have a better idea. Basically, almost all the standard fits (Gaussian, Lorentzian, etc.) can be characterized by a few basic values (mean, std for Gaussian, etc). Instead of saving a whole set of data, we can just store these as parameters of the dataset in the hdf5 file, under names that make sense (maybe fit_gaus_mean or something like that). We then can have a checkbox to turn the plotting of the fit on or off. It is probably easiest to run the fitting algorithms as a plugin.